Project Summary/Abstract Determining the computational units of a neuronal circuit is essential in understanding the function of the circuit as a whole. In many circuits, subcellular compartmentalization allows synaptic boutons on the presynaptic side, and dendrites, on the post-synaptic side, to functional independently. In sensory systems, functional divergence allows enabled the representation of many features of a stimulus using a limited number of neurons. The retina is an advantageous model system for studying synaptic mechanism of circuit motifs, like functional divergence, because our advanced knowledge of cell types and circuit architecture in the retina allows us to ask precise questions about connectivity and function in response to visual stimuli. In the mouse retina, signals diverge from two cone types to ~15 bipolar cell (BC) types, and then to > 40 retinal ganglion cell (RGC) types at the output level. Mechanisms for functional divergence are well established at the first synapse in the retina, but much less is known about mechanisms at the second synapse, from BCs to RGCs. My work will fill this gap by establishing a model system in which a single BC type diverges to provide functionally distinct input to two different RGC types. The core hypothesis of this project is that different synapses from the same BC can act as independent computational units, relaying different patterns of glutamate release to their postsynaptic partners. In Aim 1, we will obtain physiological evidence that these two RGC types share input from different synapses on the same BCs. In Aim 2, we will measure the ultrastructure of the synapses with electron microscopy to determine the role of postsynaptic RGC identity on the structure of the BC synapses. In Aim 3, we will demonstrate functional compartmentalization within the BC terminal, first with an electrotonic modeling approach, and then with experimental measurements of glutamate release. The successful completion of this project will resolve the question of whether individual BCs can transmit functionally distinct glutamate signals from different synapses. Confirmation of functional divergence at the level of BC synapses will alter our picture of the computational structure of retinal circuits. Instead of considering bipolar cells as the computational units of excitation, we will instead show that bipolar cell output synapses can perform independent computations. This paradigm change will have important consequences for models of retinal circuits, and it will open up new areas of study into the molecular determinants of subcellular wiring specificity in development.